Hacker News new | ask | show | jobs
by jonhohle 32 days ago
> But if we make a platform that is just some protocols that can't be controlled by anyone, you and I would be able to spin up anti-propaganda bots to pwn the propaganda bots without fear of repercussion.

How has this worked out with email, text messages, or the phone system, or even postal mail.

I rarely receive messages from kindly anti-propaganda bots, but sure receive a lot of messages from actual propaganda that bypass filters and infect everything like cockroaches.

Assuming that otherwise won’t happen is a basic failure to understand humanity. Spend a few hours with middle school boys and after observing their behavior, try to determine if your protocols will withstand that goofiness, naivety, rudeness, absurdness, sensitivity, callousness, puerileness, unpredictability, and rambunctiousness.

As a parent to several, I see how educational institutions (school) whose job it is to be experts at this exact behavior are failing catastrophically by not understanding this very basic idea. If your protocol something that is designed for well meaning people with good behavior who trust one another, it probably won’t work to well when given to middle schoolers and will work even worse when someone with the slightest bit of malice gets a hold of it.

1 comments

> How has this worked out with email, text messages, or the phone system, or even postal mail.

Those are centrally controlled systems where propangandists have home field advantage (email is debatable, it's halfway, it wasn't designed with the existence of companies like Google in mind). But even if that wasn't the case, it's not the same phenomenon as bots on social media. The important difference is that on social media, if there is no central moderation, the bots will cancel out eachother's influence. If I make an anti-propaganda email bot, it doesn't lower the ranking of the propaganda that's already in your inbox. But if I have an upvoting bot for their downvoting bot, they neutralize eachother.

Also, ensuring that nobody except the participants of group chats and DMs can figure out eachother's real identity is already a massive win. That alone makes it a lot harder to beat a population into submission.